Sound files 1.1. Sampling

When a sound signal from a microphone or on a tape needs to be read into a computer, it is digitized, which means that it is sampled and quantized.

Sampling is the discretization of the time domain of the signal: each second of the signal is divided up into 11025, 22050, or 44100 slices (or any other suitable number), and a sample value is associated with each slice. For instance, a continuous 377-Hz sine wave is expressed by

x(t) = sin (2π 377 t)

If the sampling frequency (or sample rate) is 44100 Hz, this sine wave will be sampled at points spaced Δt = 1/44100 second apart, and the sample values will be

xi = sin (2π 377 (t0 + i Δt))

where t0 is the time after which sampling begins. Δt is called the sample period.

Quantization is handled in the next section (§1.2).

Links to this page


© ppgb, March 30, 2004